Morphological recognition of blob shaped particles

In this paper, we proposed a morphological recognition algorithm using threshold linear superposition theory to analyze the distribution of randomly spaced and oriented blob shaped particles. In the first step, particle region is extracted from background region for fast processing. Conditional skeleton is also used to separate the overlapped particles. In the last step, we recognize the size and position of randomly shaped particles by using the hit/miss transform. Our experiments using images acquired in the texture photographic album have shown that the recognition of size and position is very effective.